Free Access
Volume 564, April 2014
Article Number A40
Number of page(s) 35
Section Stellar atmospheres
Published online 02 April 2014

Online material

Table 2

AAA-rated spectral classifications and supporting data for 213 VFTS O stars.

Appendix A: BBB classifications

The appendix addresses 139 objects in the VFTS O–B0 sample with lower-rated (“BBB”) spectral classifications, which are largely excluded from the discussion in the main text as well as entirely from the HRDs. They are listed in Table A.1. The reasons for the lower ratings are discussed next. These spectra are not displayed in this paper (except for VFTS 318 in Fig. 23) but will be at the VFTS website.

Appendix A.1: Reasons for lower-rated classifications

There are several distinct reasons for the BBB classifications. The most obvious are relatively low S/N of the data and/or severe nebular contamination at the He i lines, which render the classifications uncertain and frequently preclude luminosity classes. Of course, these cases are generally among the fainter stars in the sample. Table A.1 also contains 10 B0 spectra to define the O–B boundary; some of them are of fine quality.

A second BBB category comprises the majority of the prominent SB2. While the data quality for these is often excellent, they cannot be included in the HRDs without full orbital analyses, which in general require further observations that are in progress. (A few SB2 that were recognized later are included in the AAA list and HRDs, as are of course any that remain undetected; in general, they may be expected to have relatively fainter and less contaminating secondaries.) As already noted, the SB2 have been classified by JMA with MGB using the best data near quadratures; while some of these classifications are necessarily uncertain, others are fine.

The most vexing BBB category is discussed separately in more detail in the following section.

Appendix A.2: He ii λ4686 vs. Si iv luminosity criterion discrepancies at late-O types

Thirty-nine of the 139 BBB spectra fall into this category, as do 14 AAA types discussed in Sect. 4.1; other late-O BBB also have weak Si iv but luminosity classes could not be derived. While the classical MK procedure in general relied upon subjective averages over (sometimes mildly conflicting) multiple criteria, the tendency in this work (as well as in GOSSS; Sota et al. 2011) is to define a primary criterion for each subcategory, in the interests of clarity and reproducibility. In the case of luminosity classification at late-O types, the He ii λ4686/He i λ4713 luminosity criterion, which decreases with increasing luminosity due to emission filling in the He ii absorption line, has been preferred over the Si iv/He i ratios because of the susceptibility of the latter to metallicity effects. Rather severe discrepancies between the two criteria are encountered in many of the 30 Dor spectra at these types (as well as less frequently in the Galaxy).

Thus, this approach has unfortunately not been entirely successful in the current sample. While the different criteria agree for many late-O types in the AAA list, for a few as well as for numerous BBB they do not. A contributing factor may well be that the late-O horizontal classification criterion Si iii λ4552/He ii λ4541, the unit value of which defines type O9.7, is also sensitive to metallicity. This dependence, in combination with the rapid decline of the luminosity-dependent He ii λ4686 emission filling with advancing type in this range, such that He ii λ4686/He i λ4713 has similar values near unity at O9.7 II and B0 IV, entails substantial interrelated uncertainties

in both dimensions. However, additional sources of He ii λ4686 emission unrelated to luminosity in this young LMC sample cannot be excluded a priori. It also appears that blended multiples may be another source of this problem; several cases in which spectral classifications from HST data yielded lower luminosity classes are noted in the Table A.1 Comments (VFTS 141, 150, 153, 389). It is likely that quantitative spectral analysis and in some cases observations with higher spatial resolution will be required to fully elucidate these issues.

To quantify this problem further, the average absolute visual magnitudes of the 12 AAA and 24 BBB stars in this category with parameters are compared to the predictions from their luminosity classes here. The observed MV’s available for the BBB objects are listed in the Comments field of Table A.1; as for the discrepant AAA, the observed values are fainter than the calibration predictions (Walborn 1973) in all but one case (VFTS 389, ironically one of those with a fainter HST luminosity class). For 25 stars classified O9.5–O9.7 III, the average observed value is −3.86 ± 0.10 (m.e.), compared with a calibration value of −5.3. For six O9.5–O9.7 II-III and two O9 III, the observed value is −4.34 ± 0.12, whereas the calibration values are −5.6. There is one O9.5 II with an observed −4.6 but calibration −5.9; and one O9.5 IV with values of −3.8 vs. −4.7, respectively, but another (VFTS 389) with a brighter observed value of −5.1. The observed values do increase toward brighter luminosity classes, but except for the two class IV stars, the observed minus predicted are remarkably consistent at 1.3–1.4 mag, as also with the O9.5–O9.7 V calibration value of −4.1 and what would have been predicted from the small Si iv/He i ratios. On the other hand, for three B0 V stars (VFTS 347, 496, 540) the observed values are more consistent with the calibration, average −3.5 vs. −3.6 (Walborn 1972), respectively, so that some of the apparent late-O giants could also be slightly cooler objects as discussed in the previous paragraph.

Appendix A.3: Multiplicity

Sana et al. (2013) have determined from detailed radial-velocity analysis that 35% of the VFTS O stars are spectroscopic binaries with amplitudes greater than or equal to 20 km s-1, corresponding to an intrinsic (bias-corrected) binary fraction of 51% for the orbital-period range considered (P < 103.5 days). They also detect an additional 11% with smaller amplitudes, some of which may have other physical origins, but all of which are characterized as “SB” in Tables 2 and A.1 here for simplicity, with the amplitude ranges specified. All of the SB designations here are from the work of Sana et al., except for the SB? (with displacements between stellar and nebular lines but no significant variations detected) which have been added here. In addition, visual multiples have been so designated from inspection of HST/WFC3 images here (but note that there was no WFC3 coverage at the time for some objects at the peripheries of the VFTS field). As specified in the notes to the tables, all of these categories sum to similar total multiplicity rates of 140/213 or 66% for the AAA spectral types (Table 2), and 96/139 or 69% for the BBB (Table A.1). As already noted, most of the SB2 are in the BBB list, while the AAA SB are predominantly SB1. As also pointed out in the table notes, these percentages are lower limits because of the detectability gap between radial velocities and direct imaging, as well as unfavorable inclinations in the case of the former. The implications for astrophysical and population studies of O stars in the Magellanic Clouds and beyond are clear and should always be borne in mind.

Table A.1

BBB-Rated spectral classifications and supporting data for 139 VFTS O–B0 stars.

© ESO, 2014

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.